DevOps of Service Assets for Telco Cloud Automation

Size: px
Start display at page:

Download "DevOps of Service Assets for Telco Cloud Automation"

Transcription

1 DevOps of Assets for Telco Cloud Automation Copyright Huawei Technologies Co., Ltd All rights reserved. No part of this document may be reproduced or transmitted in any form or by any means without prior written consent of Huawei Technologies Co., Ltd. Trademark Notice, HUAWEI, and are trademarks or registered trademarks of Huawei Technologies Co., Ltd. Other trademarks, product, service and company names mentioned are the property of their respective owners. General Disclaimer The information in this document may contain predictive statements including, without limitation, statements regarding the future financial and operating results, future product portfolio, new technology, etc. There are a number of factors that could cause actual results and developments to differ materially from those expressed or implied in the predictive statements. Therefore, such information is provided for reference purpose only and constitutes neither an offer nor an acceptance. Huawei may change the information at any time without notice. HUAWEI TECHNOLOGIES CO.,LTD. Huawei Industrial Base Bantian Longgang Shenzhen ,P.R.China Tel: HUAWEI TECHNOLOGIES CO., LTD.

2 The Crafting Center solution The DevOps automation platform: Agile & Intelligent Design and Orchestration (AIDO) transformation for continous service development and operations The zero-touch operation system: Intelligent Engine System (IES) Model-driven service creation for agile service assets development Professional service supported by the cross-functional agile team What is a Crafting Center? The SCC architecture Requirements Management Asset Design & Creation Testing Asset Repository Operation Runtime The 3 core competencies of Crafting Center Overview Supporting 5G and network as a service Building service development ecosystem Optimizing zero-touch operations by AI/ML API-driven, open platform DevOps centric processes Cross-functional agile team of people

3 Executive Summary Development and Operations is the Heart of Digital Transformation With the competition from over-the-top (OTT) service providers, Communication Providers (CSP) are under pressure to provide better services in a quicker and cost effective manner. In recent years, many CSP have started their digital transformation program to transform themselves to become a Digital Providers (DSP) aiming to provide better service to their customers in a more agile manner. One of the key success factors achieving this is for a CSP to transform the service lifecycle management process from how a service is created, managed and operated. Having a competency center within a CSP called the Crafting Center (SCC) can be setup to be responsible for this purpose. As for any organization, SCC should be have a team of experts, processes and an automation platform in order to have capabilities to create and manage services in an agile matter, supporting the CSP transforming into a DSP. In this whitepaper, Huawei will propose the SCC architecture and the required competency of a SCC and share our global experience with our CSP customers on how SCC can help them during their digital transformation journey. 2.1 transformation for continous service development and operations Digital transformation is the hottest topic recent years in the telecommunications industry. A digital transformation program is a continuous effort that services can be designed, provisioned, healed, and optimized very easily and automatically. This goal cannot be achieved without transforming how services are being created, managed and operated today. With the latest technologies, such as Network Functions Virtualization (NFV) and Software-defined networking (SDN), CSPs can realize the promise of making networks become more flexible and scalable so that new services can be deployed quickly and efficiently, enabling service self-provisioning and network self-healing to become achievable. In order to achieve this goal, the current way of creating services must be changed. It usually takes months to make a new service commercially available as it involves many departments and internal processes. Even a mature service, such as Virtual Private Network (VPN) for an enterprise, can sometimes take more than 4 weeks to prepare before a customer can use it. It will take even longer time if there is customization of the service that is requested. Customer care, network operation and account management will also need to be involved to handle any issues that may arise such as service outage and quality of service are breached. transformation must be implemented to overcome these challenges

4 Crafting Center: DevOps of Assets 2.2 Model-driven service creation for agile service assets development 3.1 What is a Crafting Center? To achieve service self-provisioning and network self-healing, CSPs must transform the existing way of how services are created, managed and operated. In the constant changing environment, different customers will have different requirement about a specific service, such as add-on services, quality of service and change requests. More often than not, when a new service idea comes from marketing or customer requirements, the CSP cannot respond fast enough because of the complexity of its current internal service creation processes. The process involves many different functional departments such as planning, engineering and operation teams, where work details are also not visualized from team to team. A model-driven approach, using service assets to streamline and create visibility of a service at every stage must be applied to solve these issues. All service components should be modeled for easy access and fast service creation. assets should have different categories include service models, network functions specifications and interfaces, operation policies and event handlings. With the availability of these service assets, a service master or service creator can utilize these assets to compose and define a service that is required by the market or a specific customer. This process should be done with minor configurations of the service assets and without software coding. Once the service is created containing all service behaviors and operation policies, it can then be sent to the testing environment for validation and finally to the production environment for the actual deployment of the service. The CSP s goal of service transformation is to create different service assets using the model-driven approach so that a service can be created quickly, updated regularly with continuous integration, continuous delivery, testing and operated in a zero-touch fashion. 3.2 As described in the previous section, creating service assets in a continuous manner is the key success factor of service transformation. A competency center, called Crafting Center or SCC, should be setup within the CSP to develop, manage and operate service assets. The key responsibility of SCC will be handling service requests, create service assets and services, testing services and analyzing services for service asset improvement after the services are running in the operation mode. In other words, SCC is the competency center for continuous development, continuous integration and continuous deployment of service assets, or the DevOps of service assets. The SCC architecture For continuous development, continuous integration (CI), continuous testing (CT) and continuous delivery (CD), continuous deployment (CD) of service assets, the SCC architecture should have four basic functions. They are: i. Requirements Management ii. Asset Design & Creation iii. Testing iv. Asset Repository v. Operation Runtime As described below in Figure 3-1, the 5 functional blocks defined how they work together and how service assets will be developed, managed and operated. These five functional blocks formed the closed-loop lifecycle and DevOps process of service assets

5 Crafting Center (SCC) 1 Requirements Management 2 Design AI Cross-functional Agile Team Asset Design & Creation 4 Portal Plan Asset Repository s Operation Testing Build & Test Delivery Resources Operation Runtime Orchestration Self-optimization Assurance Data Collection, Analytics & Events (DCAE) 5 Selfprovisioning Selfhealing Asset Design & Creation The Asset Design & Creation is responsible to design and develop services that are defined in Requirements Management. A graphical and simple to design interface or portal should be available to Asset Design & Creation for both the Customer Facing and Resource Facing design and creation. A typical service, such as a VPN, is composed of several component such as universal gateway, firewall and security. This Asset Design & Creation function will be responsible for both creating all the required service components and glue different service component together to form the actual service that can be offered to customers. Ideally, service components with all the necessary parameters and interfaces should be created once and stored in the central database that allow different services to (re)use these service component to form a desired service. Major services will also be stored in the same fashion for easy extraction and to allow minor customization for different customer needs Testbed Figure 3-1 Crafting Center (SCC) Architecture Requirements Management Network Requirements Management is responsible to analyze service requests and kick start the service creation process. The requests can be within CSP s sales and marketing teams or directly from the end-customers. The service requests can be described in the context of minimum viable product (MVP) which provides the service features and minimal desirable experience (MDE) that specify the customer experience and service quality attributes of the service. The Requirements Management function will also be responsible for any suggestions for improvements or adjustments after the services are launched in the production environment It is critical to take a model-driven, drag-and-drop approach for service creation as it allows agile service development without software coding or customization. Every service asset should be modeled for its service features, interconnection requirements, resources requirements and operational requirements. Asset Design & Creation should have the capabilities for different domain experts such as wireless, core, transmission and networks to create their respective service asset using the same user interface and portal. The portal can also be made available for 3rd party developers if the CSP choses to open this capability to their partners in their ecosystem to develop different service assets. Testing Once a service is created by Asset Design & Creation, it is required to be tested and checked if the service can satisfy requirements such as functionality, performance, reliability and security. Testing should use the test catalogue approach executed across four dimensions: 05 06

6 Network Function & Network, Test domain and lifecycle dimensions together specify the test object and context of the test catalogue with Automation & Agility dimension specifying the correlated executable test artifacts and runtimes which support the testing automation process. The Testing function maintains a repository of test catalogues, shielding the technology complexity existing from either the test object or test context thereby enabling the fast execution of the specific validation testing tasks Asset Repository Asset Repository is the central database that stores all service assets. This is the most important part of SCC as it keeps the knowledge of all services being created. The repository will store service assets including service models, resources components from various domains including wireless, core, transmission and networks, operation use cases, and policies and testing cases. The Asset Repository function will also be responsible for distributing the service in to runtime environment for service orchestration and assurance after the service has been created and tested. The service asset in the repository will be continuously refined and updated by various domain experts to keep current with the latest customer and service requirements. 3.3 The 3 core competencies of Crafting Center The key success of a Crafting Center is to have a rich database of service assets and the SCC architecture has defined the necessary functions required to create and manage service assets in an agile matter. To implement and execute these functions, three major competencies are required: i. API-driven, open platform ii. DevOps centric processes iii. Cross-functional agile team of people An open, API-driven platform for fast service asset creation facilitating processes for continuous development of service by teams of experts from all functions including planning, design, build, test and operation in all domains such as wireless, core, transmission and network. Figure 3-2 below illustrates such dependencies. Cross-functional Agile Team Assets Operation Runtime The Operation Runtime is responsbile for the execution and operation of service assets in the production environment. To drive for zero-touch operations, Operation Runtime function should have 3 critical automation proccesses including self-provisioning, self-healing and self-optimization to create, run, monitor, analyze and optimize the service that is defined in Requirements Management. The keys modules of Operation Runtime include Orchestration, Resource Orchestration and Data Collection, Analytics, and Events (DCAE), where is responsible for preparing and orchestrating all necessary services and resources of a service, monitoring and analyzing service behavior for network service and quality optimization. 07 DevOps API-driven,Open PLATFORM Figure 3-2 PROCESS PEOPLE The Core Competencies of Crafting Center 08

7 3.3.1 API-driven, open platform An Application Programming Interface (API) driven, open platform system is required to execute all functions defined in the SCC architecture. An API-driven design creates the standardization of functions and interfaces between functions. APIs will only expose certain characteristics of a function, allowing greater control and minimizing risks of a function. Another benefit of an API-driven platform is that, the implementation of an individual function can be updated, changed or replaced without affecting the behavior of a function because the behavior of function is being standardized by its APIs. Some functions will utilize open source implementation, such as Open Network Automation Platform (ONAP), while other functions will be implemented by experts or vendors to perform more sophisticated and complicated tasks such as artificial intelligence (AI) and machine learning (ML). The design principle of the platform is that it should be open for easy function plug-in, open interfaces by APIs for accessing different function capabilities. The platform should also support the complete lifecycle management functions including planning, service & network design, building, testing, capabilities to chain and automate these functions and a central repository for service assets management and distribution service corrections, modifications and refinement. The DevOps methodology with the continuous integration/continuous delivery mechanism is critical for having a closed loop and automation operation of the entire service creation and management lifecycle. Cross-functional agile team of people The cross-functional agile team will be a team of people to use the platform and process and carry out all functions defined in the SCC architecture. The team should have different experts including Domain Manager/Scrum Master, Design & Creation Architect, Creation Engineer, Modelling & Algorithm Engineer, Software & Platform Architect and Software DevOps Engineer to carry out different functions: Domain Manager/Scrum Master is responsible for the overall delivery and experience of a specific service domain. Domains such as video, smart home and smart city. Design & Creation Architect is responsible for network architecture design for realization and implementation of services DevOps centric processes The service creation process needs to be automated to scale with the dynamic and ever changing business environment is important for a CSP. s should be refined and updated according to market or customer requirement changes. With this continuous effort of service creation and operation, the DevOps methodology must be applied to be successful. The service creation and operation tasks should be seamlessly chained together for greater efficiency. Tasks including planning the services, design the services, creating the service assets including service model, operation events and policy, testing cases, testing the services in a simulated environment or the test bed, registering the service in the central repository and distributing the service in the runtime environment. The service creation process will repeat itself for Creation Engineer is responsible for service design and orchestration of a specific technical domain. Domains include wireless, transmission, core, and IT. Modelling & Algorithm Engineer is responsible for resources, services, and policies modelling, AI algorithm design and big data analytics. Software & Platform Architect is responsible for software architecture and platform design in order to support agile asset modeling, creation and service development. Software DevOps Engineer is responsible for software programming, testing and verification

8 Huawei's Crafting Center Solution The DevOps automation platform: Agile & Intelligent Design and Orchestration (AIDO) 4.1 Professional s The Crafting Center Solution Huawei s Crafting Center Solution is designed to help a CSP to setup, implement and run the SCC. The Huawei Crafting Center Solution composed of the automation platform Agile & Intelligent Design and Orchestration (AIDO) platform, operation system Intelligent Engine System (IES) and professional services supported by the cross-functional agile team. Network Integration IES & Integration Design & Orchestration Testing Continuous Operation Automation The Agile & Intelligent Design and Orchestration (AIDO) platform is responsible for the development, testing and storage of service assets: Design Studio is the graphical interface to allow users to develop service assets without software coding. Users can extract services from asset repository for update or modification or create a service from scratch. Planning & Design, Automation Testing & Integration is the integration tool suites that are responsible for multi-vendor integration which includes plan, design, deploy and testing. The service creation and operation processes define the business logic of a task and these tools are chained together to perform this task automatically. Different services can have different processes and therefore the chaining logic of these integration tools can be different. MDE MVP Data Analy cs & AI MDE: Minimal Desirable Experience MVP: Minimum Viable Product Figure 4-1 Design Studio Planning & Design Automa on Tes ng & Integra on Assets Repository s Resources Opera on Tes ng Cross-functional Agile Team Orchestration Orchevra on Resource Orchestra on Optimization Huawei Crafting Center Solution Dynamic Inventory Management Data collec on & analy cs Assurance Policy Analy cs & Execu on Assurance & Performance Management Policies & rules op miza on Asset Repository will be responsible for storing, distributing of all services assets. assets include service models, testing use cases, operation policies and event handling cases. It act as the central database for all service asset created and tested by the Design Studio and Testing. It also be responsible for distribution the service to the production environment for service orchestration and assurance. The different service assets can be only accessed by different authorized users. AIDO is designed based on the architecture specified in section 3.2 with the following important characteristics: a.devops model for continuous design, testing and delivery. AIDO adopts the DevOps model for service asset development, including activities such as planning, design, build, testing and delivery. This provides an agile environment for continuous development of service assets. b.rich service assets. With our experience and global practices, AIDO Asset Repository already has built a repository of many different service assets 11 12

9 will manage these workflow and tasks at all stages. Different tasks can be assigned to different users according to their skills and capabilities. f. Compliance with ONAP design time. AIDO is compliant with ONAP design time. AIDO has also implemented additional features to support DevOps and full life cycle management, as described above. AIDO is committed to compliance with future release ONAP to maintain consistency and latest features from the open source community. that can be made available for other customers to (re)use. asset types in the repository include service models, resources, testing and operations. c.ai engine for best network design. Network design is a complex and professional work as it requires expertise from translating service requirement and quality into network requirement and topology design. The network design for each CSP will also be different as the network topology and attributes of every network is different. The AIDO AI engine will generate the best possible network design based on the hundreds of network topologies from our global experiences, service requirement, quality expectation, and current network situation. Network automation cannot not be achieved without an operation platform to manage automation processes for service creation and management. This is exactly what AIDO is targeted to do. AIDO will be offered by the as-a- model. This provides a flexible business model for customers and easier access to the fundamental functions required for the SCC. The AIDO model also allows customers for faster access of the latest functions of AIDO and service assets from Huawei s global project experiences. AIDO can also be seen as the true implementation of the CSP s DevOps service assets model. d.open ecosystem supported by open API. The functions and service assets of AIDO can be accessed through a set of open API. This allow CSP, their partners and other 3rd party service developers using the AIDO standard API programming for agile service asset development. e.full life cycle management. assets are required to be managed during all stages of development that includes planning, design, build, testing and delivery. The workflow and tasks for developing different services can be defined using AIDO. Once the workflow and tasks are defined, AIDO Figure 4-2 The AIDO automation platform 13 14

10 4.1.2 The zero-touch operation system: Intelligent Engine System (IES) Intelligent Engine System (IES) is the operation system responsible for service orchestration, assurance and optimization. IES will be installed in the production environment for operating CSP s live service and network. It takes the service assets generated by the AIDO asset repository for processing, monitor the services for assuring service Key Performance Indicator (KPI) and provide service analytics for further service enhancement. Zero-touch operations is the ultimate goal of IES by fully automating all operation processes and activities. To support zero-touch operations, it has the following features: a.3 automation processes. The 3 automation processes include selfprovisioning, self-healing and self-optimization. Self-provisioning will be conducted by Orchestration module where it orchestrates and prepares service and network configuration for production use. Self-healing is the capability where the Assurance module monitors service behavior and KPI in order to trigger actions such as scale-in/scale-out resources, root cause analysis and service automatic recovery when it is needed. Selfoptimization will optimize the services by constantly enhancing the scaling policies, performance recovery rules. Self-optimization will also provide information for AIDO to optimize the AI engine and service design. d. Compliance with ONAP run time. Similar to AIDO, IES is compliance with the ONAP run time. IES is committed to compliance with future release ONAP to maintain consistency and latest features from the open source community. As CSPs open up their network and provide 3rd party service providers the capabilities to develop different services through open API, an intelligent operation system that can run, monitor and optimize service and network automatically is required, as it is not possible for CSP manage the services and networks by using the traditional reactive and manual driven approach. IES is positioned to help CSP to transform towards zero-touch operation using the automated and standardized design approach. b.ies, the next generation of OSS. IES is the next generation of OSS where it is responsible for service and resource orchestration, service and network assurance. With the 3 automation processes, IES aims to achieve zerotouch operations, helping CSP to simply the operation of future services and networks. c.hybrid network support. IES is capable of management both cloud and legacy network in the hybrid environment. CSP will only need to worry about the overall service quality of service and network KPI, leaving the network details underneath for IES to manage. Open API will be used to implement the operation functions, making the functions to be standardized, consistence and future-proofed Professional service supported by the cross-functional agile team Our professional service, supported by the cross-functional agile team with experts from different functions, will help CSPs to create, manage and operate services in the SCC. Professional services include service design and orchestration service, continuous operation automation service and testing service. Our cross-functional agile team brings in years of experience from various domains and functions, service transformation experience to help CSPs adopt the SCC mechanism without large investment on human resources and training programs

11 Challenges Long Time Readiness 4 weeks Many Tasks Multi- Depart ments Case Study of Huawei Crafting Center Solution Tasks Stage1 0-1 Stage 2 1-N Planning 5 weeks 2 weeks 2 days 1 2 VPN Design PS APN Design Crafting Center WAN Design NFVI Design Planning & Design processes modeling Planning & Design adjustment 3 Logic Design VNF/Resources Orchestration Development Asset Repository Orchestration Development Interfaces, deployment, configuration modeling Interfaces, deployment, configuration models adjustment and models replication Testing Publishing Execution Testing processes modeling Testing models adjustment Runtime Execution Execution Execution Huawei s Crafting Center Solution has helped to automate this process based on the AIDO automation platform: i. From 0 to 1 First, by applying the model-driven approach, all the service asset models including planning, design, service and network orchestration templates will be developed. Then the integration, provisioning, testing, orchestration processes are created with developed assets by orchestration them together. AIDO has built the foundation of the Mobile VPN service and published in service assets repository. ii. From 1 to n Different enterprises will request different form of the Mobile VPN service. Different requirement such as higher security, dynamic bandwidth and better quality of service. The service master will process the service request by extracting the service assets from the repository by applying minor configurations and updates. As the result, a specific mobile VPN service for a specific customer will be ready in 2 days, significantly reduced compared to the original time required. Figure 5-1 A Case Study of Huawei Crafting Center A leading CSP in Asia has a digital transformation program aiming to transform themselves to be more agile in response to customer and market needs. As one of the largest CSP in the local market, this CSP has a large base of enterprise customers and it is very important for this CSP to be innovative and responsive to them. The Mobile VPN service. An important goal for the digital transformation program is transform how services are created and managed. The Mobile VPN service is used as the first service being created and managed using the SCC architecture. The current Mobile VPN service offered to their enterprise customers is taking too long to be ready, taking more than 4 weeks to complete from receiving the order from the customer to preparing the resources and operations. 6.1 The Evolving Crafting Center Solution Overview Huawei s Crafting Center Solution helps CSPs to transform service creation and management using the automation platform AIDO, operation system IES, supported by our professional service team. As 5G is around the corner, the Crafting Center Solution is already planned to support 5G, AI/ML for zero-touch operations and building up an open ecosystem for asset development

12 6.4 Optimizing zero-touch operations by AI/ML SCC Crafting Center AIDO IES AI / ML Asset Ecosystem 5G / Network as a Operation automation is important as more and more services can be offered by the SCC since manual network and service operation practices and policies cannot cope with the diversity and constant change of network and services. To achieve operation automation, AI and ML is the critical technology required for preventive maintenance, network and service self-healing and selfoptimization. Our deep experience on both traditional and NFV/SDN based network provides a solid foundation and vast amount of data to develop operation artificial intelligence algorithms that train the machine learning engine. The operation rules and policies will be constantly updated based on the knowledge learned from our global network project experience. Figure 6-1 Continuous Evolution of Crafting Center Supporting 5G and network as a service 5G promises a greater network flexibility and business opportunities for CSPs to provide services such as network slicing and network as a service. The key for offering these types of services is to open up the network for 3rd party developers to create services. Huawei s Crafting Center Solution framework design is already supporting this and will be available to the market in the coming release. Building service development ecosystem Having a rich database of service assets is vital as they are key ingredients of creating services within the SCC. An open ecosystem allowing 3rd party providers to develop service assets will greatly improve CSP s competitive and responsiveness to market needs. Therefore, AIDO is important to build an ecosystem and enrich the service asset repository. As the AIDO platform is available in the Cloud, CSP partners from different domains and expert areas can create, test and run their service assets on AIDO, supporting CSP for agile service development. Conclusion Digital Transformation requires structural changes to the CSP s business strategy, models, operations, products, marketing approach, and objectives by adopting digital technologies, aiming to accelerate sales and growth of the business. An environment for CSPs to create and manage services in an agile and automated way is vital to the success of digital transformation. Crafting Center with the automation platform, processes and professional experts should be setup in a CSP to provide such an environment. The roads of digital transformation is a continuous effort and CSPs will need a strategic partner for this long-term and complex transformation journey. The Huawei s Crafting Center Solution, together with our global experience on NFV/SDN/OSS projects, have enabled us to be the perfect choice for this partnership